Advanced Analytics Framework for Omnichannel Leads Optimization

Mahadevan Jayaram, CEO at Marketsof1, talks about how one can make an astute use of the ever-changing digital environment through a well-thought advanced analytics framework.

Why an advanced analytics framework?

Online/digital sourcing of leads is now a bloody war. Google changes its algorithm at will. Consumers change their search terms without bother. Consumers’ taste and preferences change by the hour and by geography. In this ever-changing environment, competition brings in added complexity. Within organizations too a lead sourcing team, a lead management team and a lead fulfillment team largely operate in silos. Therefore, for each successful lead, organizations end up paying more than necessary and is increasingly becoming unsustainable.

This is particularly true for auto manufacturers, Banks, Insurance companies etc. where a lead hops from a digital channel to a telephone call to a home or showroom visit. Fixing this requires a set of sophisticated analytics and omnichannel analytical CRM to improve performance through the funnel. An end to end solution that improves every stage of the funnel thus improving the aggregate performance of the sales funnel significantly

A typical scenario faced by lead generation teams

Many digital campaigns are leading to –

1. Low visits to leads ratio and then
2. Low conversion (e.g. Test-drive or say Purchase) to Leads ratio

Assumptions:

1. Visits are a function of how good are the
   • Paid campaigns in Google, FB, etc.
   • SEO functions in the search engines
   • Social free campaigns
   • Remarketing campaigns

2. Leads are a function of how good are the
   • Landing page marketing content (e.g. text, image, video etc.)
   • The user experience (UX) at the initial customer touchpoint (Moment of truth)

1. Conversions are a function of
   • How good is the customer experience (CX) through the eval to purchase journey
   • Post lead call, email, sms etc.
     1. Time and context of the interaction
     2. Quality of interaction

Determination of prospective lead from a visitor:

The first step in the optimization process is to separate the wheat from the chaff. In this case, it means how to do you cull out the quality leads that are most likely to convert from the ones that would not result in business. This would typically start with analysing the clickstream data for past campaigns for a period of at least 6 months.

Clickstream data provides immensely valuable data of a visitor. Among others, basis a user id generated from a cookie, the following data can be secured, if tracked for a minimum period of six or seven months:

1. Campaign and landing page through which the visit was initiated
2. Total pages visited & Total unique pages visited
3. Time of Visit
4. Time spent on site
5. User journeys on the website
6. Other demographic data of the visitor such as a country, city, region, speed, ISP, Home/Biz etc.
7. The previous site visited
8. The keyword of the referrer search engine

Once this data is available, running it through a machine-learning algorithm provides us with insights about the variables that influence the completion of a successful lead. There are specific algorithms for this purpose, such as the Support Vector Machines (SVM) algorithm. The end result of this process is an analytical model that requires it’s hypothesis to be tested.

Testing the Hypothesis

The model which is most cases would be a small piece of code gets deployed in real-time. Which essentially means the campaign manager has more control of the variables that influences successful lead completion. For eg: The campaign handler can differentiate based on demogs between high & low propensity customers. Or based on behaviour; what is the best time for conversion? Based on the model’s suggestions, keep tweaking it till it reaches optimal performance.

Benefits of using advanced analytics in the optimization of  the leads funnel

Once the path is known for best conversions…the site is reimagined to nudge the user in the stated path. An algorithm will continuously monitor the buyer actions, across all visits, and score. Once the score reaches the threshold to convert into a lead… aggressive retargeting will be started. (It tells us who to target). Avoid campaigns on sites that do not help convert. Or avoid advertising sources that don’t matter.

It starts with experimentation (Design of Experiments – a statistical analytical method) of various combinations over a small sample of leads. This helps understand, what text, channel, creative and wave are most effective. The most successful method then is automated using an analytical CRM tool.

Final Takeaway

A well thought out advanced analytics framework works on the TOFu, MOFu, and BOFu. It removes the negative effects of siloed analytics in these three stages by combining them under one integrated Machine Learning Algorithm. The result is an exponential impact at every stage of the funnel leading to both increased leads and even more increased conversion.